The yahoo webgraph has 1.4B vertices and 8B+ edges. Stripping the hot ones
is not an option :).

Unfortunately, the vertex/page with most in-edges/links has 759599052 of
them, which is smaller than both (2^31)/4 and (2^30)/4 :). I have to find a
way to solve this.

One way to handle supernodes would be what GPS people do by replicating a
vertex into multiple replicas that contain a partition of the original
adjacency list. Maybe I could open up a JIRA for this.



On Thu, Feb 21, 2013 at 8:33 PM, Eli Reisman <[email protected]>wrote:

> Sorry thats what I meant, the supernode got the messages, and boom. So even
> using the new byte buffering, one per vertex, you are hitting the wall
> here. This is FloatWritible messages, converted to 4 bytes each, in a byte
> buffer, and the hard limit for messages per vertex is going to be (2^30)/4
> and thats that. Interesting. How many vertices/edges in this yahoo webgraph
> you're running? Stripping them out via preprocessing MR job or something is
> not an option for this job?
>
> It is a shame to have a totally processable graph at your cluster's scale
> and have a couple supernodes blow up your job anyway. That will never stop
> being a problem for us as we scale to larger and larger datasets. It would
> be fun to explore some ways of handling supernodes as a "special case"
> since they can be identified so early in the job run.
>
> With YARN, fine grained preprocessing with only a fraction of the nodes we
> would use for our Giraph workers, before starting the worker tasks or the
> job, could assess the input graph and leave hints and instructions for our
> Giraph workers. Even assigning the most well-provisioned hosts in the
> cluster by hand to handle data partitions that contain supernodes and are
> likely to be traffic-heavy is possible. But lots of things are possible
> with YARN that don't solve problems for Giraph as a whole.
>
> On Thu, Feb 21, 2013 at 7:22 AM, Maja Kabiljo <[email protected]> wrote:
>
> > Yeah, that's exactly what I told you.
> > We have one ExtendedDataOutput per vertex. And the limit is about
> > (2^30)/4, since you will have double of the size allocated at some point.
> > What do you mean you don't want to use a combiner for evaluation reasons?
> >
> > On 2/21/13 5:38 AM, "Claudio Martella" <[email protected]>
> wrote:
> >
> > >Yep. Actually more than a supernode sending messages it looks more like
> a
> > >supernode receiving a lot of messages, and hence filling the inbox queue
> > >(in bytearrayformat).
> > >I assumed a max (2^31)/4 (for float messages) limit because I assumed we
> > >had such an object per vertex, but it looks like the
> ByteArrayOutputFormat
> > >is used for more than a single vertex.
> > >Maja, do you think it could make sense to use one array per vertex (I
> > >haven't looked deep at the code, but I assume that we have one per
> > >partition now)?
> > >
> > >
> > >
> > >On Thu, Feb 21, 2013 at 12:49 AM, Eli Reisman
> > ><[email protected]>wrote:
> > >
> > >> Sorry didn't see the bottom of this thread before! I like Maja's idea,
> > >>some
> > >> mega-supernode (thats a technical term) just sent that vertex so many
> > >> messages that the byte buffer size overflowed on the right shift. I
> saw
> > >> this problem in some other forms last summer. The number of other ways
> > >>that
> > >> number could go negative seem limited.
> > >>
> > >>
> > >> On Wed, Feb 20, 2013 at 3:41 PM, Eli Reisman <
> [email protected]
> > >> >wrote:
> > >>
> > >> > I do know when you run out of memory it often looks like the logs
> are
> > >> > fine, or seem to end in a wierd spot, but really the last X number
> of
> > >>log
> > >> > messages just doesn't get logged. Sometimes its worth looking at the
> > >> worker
> > >> > logs individually and one in every group of them will actually get a
> > >> > sensible exception message logged. Don't know if thats the case
> > >>though. I
> > >> > would tend to agree with your intuition regarding the queues and
> > >> > concurrency.
> > >> >
> > >> > On Tue, Feb 19, 2013 at 3:04 PM, Claudio Martella <
> > >> > [email protected]> wrote:
> > >> >
> > >> >> Looking at the second line of the logs, it does not look like the
> > >>case,
> > >> >> but
> > >> >> it could be. do you have any suggestion for the flags in
> particular?
> > >>I'm
> > >> >> running with -XX:+HeapDumpOnOutOfMemoryError  now, and I don't get
> > >>any
> > >> >> dump.
> > >> >>
> > >> >> It could be the case, but the behavior is quite consistent
> currently,
> > >> I'd
> > >> >> get a Child error/OOM pretty soon in these cases.
> > >> >>
> > >> >>
> > >> >> On Tue, Feb 19, 2013 at 11:45 PM, Avery Ching <[email protected]>
> > >> wrote:
> > >> >>
> > >> >> > Can you add GC flags?  Perhaps you are running into full GCs...
> > >> >> >
> > >> >> >
> > >> >> > On 2/19/13 1:45 PM, Claudio Martella wrote:
> > >> >> >
> > >> >> >> I have not investigated further yet, but this does not look
> > >>connected
> > >> >> to
> > >> >> >> ZK, but mostly to some concurrency on the queues. The number of
> > >> >> requests
> > >> >> >> stays to 29...
> > >> >> >>
> > >> >> >>
> > >> >> >> On Tue, Feb 19, 2013 at 9:58 PM, Sebastian Schelter
> > >><[email protected]>
> > >> >> >> wrote:
> > >> >> >>
> > >> >> >>  For me, Giraph on Hadoop 1.0.4 runs into some kind of deadlock
> > >>after
> > >> >> the
> > >> >> >>> input was loaded when using it with 200 workers on a
> > >>billion-scale
> > >> >> >>> graph....
> > >> >> >>>
> > >> >> >>> I tried several versions of zookeeper, but nothing helped.
> > >> >> >>>
> > >> >> >>> Best,
> > >> >> >>> Sebastian
> > >> >> >>>
> > >> >> >>>
> > >> >> >>> On 19.02.2013 21:55, Claudio Martella wrote:
> > >> >> >>>
> > >> >> >>>> I'm running PR with trunk on yahoo webgraph across 528 workers
> > >>and
> > >> I
> > >> >> get
> > >> >> >>>> the workers hanging at the flushing:
> > >> >> >>>>
> > >> >> >>>> 2013-02-19 21:40:07,187 INFO
> > >> >> org.apache.giraph.graph.**ComputeCallable:
> > >> >> >>>>
> > >> >> >>> call:
> > >> >> >>>
> > >> >> >>>> Computation took 28.386961 secs for 95 partitions on superstep
> > >>0.
> > >> >> >>>>
> > >> >> >>>   Flushing
> > >> >> >>>
> > >> >> >>>> started
> > >> >> >>>> 2013-02-19 21:40:07,604 INFO org.apache.giraph.worker.**
> > >> >> >>>> BspServiceWorker:
> > >> >> >>>> finishSuperstep: Waiting on all requests, superstep 0 Memory
> > >> >> >>>> (free/total/max) = 5993.42M / 8661.44M / 14346.69M
> > >> >> >>>> 2013-02-19 21:40:07,605 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 558 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 6.4758,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0003, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 27.083
> > >> >> >>>> 2013-02-19 21:40:26,717 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 39 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 3.7965,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0.0001, MBytesSent = 175.3834, MBytesReceived = 0.0068, ave
> sent
> > >> req
> > >> >> >>>>
> > >> >> >>> MBytes
> > >> >> >>>
> > >> >> >>>> = 0.2998, ave received req MBytes = 0, secs waited = 46.195
> > >> >> >>>> 2013-02-19 21:40:32,248 INFO
> > >> >> >>>> org.apache.giraph.comm.netty.**handler.RequestDecoder: decode:
> > >> Server
> > >> >> >>>>
> > >> >> >>> window
> > >> >> >>>
> > >> >> >>>> metrics MBytes/sec sent = 0.0002, MBytes/sec received =
> 3.9824,
> > >> >> >>>>
> > >> >> >>> MBytesSent
> > >> >> >>>
> > >> >> >>>> = 0.0066, MBytesReceived = 155.4787, ave sent req MBytes = 0,
> > >>ave
> > >> >> >>>>
> > >> >> >>> received
> > >> >> >>>
> > >> >> >>>> req MBytes = 0.0155, secs waited = 39.04
> > >> >> >>>> 2013-02-19 21:40:42,041 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 36 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 2.8509,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0.0001, MBytesSent = 175.3834, MBytesReceived = 0.0068, ave
> sent
> > >> req
> > >> >> >>>>
> > >> >> >>> MBytes
> > >> >> >>>
> > >> >> >>>> = 0.2998, ave received req MBytes = 0, secs waited = 61.519
> > >> >> >>>> 2013-02-19 21:41:07,456 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 2.0174,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0.0001, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave
> sent
> > >> req
> > >> >> >>>>
> > >> >> >>> MBytes
> > >> >> >>>
> > >> >> >>>> = 0.2998, ave received req MBytes = 0, secs waited = 86.933
> > >> >> >>>> 2013-02-19 21:41:22,456 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 1.7205,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0.0001, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave
> sent
> > >> req
> > >> >> >>>>
> > >> >> >>> MBytes
> > >> >> >>>
> > >> >> >>>> = 0.2998, ave received req MBytes = 0, secs waited = 101.934
> > >> >> >>>> 2013-02-19 21:41:37,457 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 1.4998,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0.0001, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave
> sent
> > >> req
> > >> >> >>>>
> > >> >> >>> MBytes
> > >> >> >>>
> > >> >> >>>> = 0.2998, ave received req MBytes = 0, secs waited = 116.935
> > >> >> >>>> 2013-02-19 21:41:52,458 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 1.3293,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0.0001, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave
> sent
> > >> req
> > >> >> >>>>
> > >> >> >>> MBytes
> > >> >> >>>
> > >> >> >>>> = 0.2998, ave received req MBytes = 0, secs waited = 131.936
> > >> >> >>>> 2013-02-19 21:42:22,459 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 1.083, MBytes/sec
> > >> >> received
> > >> >> >>>> =
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 161.937
> > >> >> >>>> 2013-02-19 21:42:37,460 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.9912,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 176.938
> > >> >> >>>> 2013-02-19 21:43:07,461 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.8475,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 206.939
> > >> >> >>>> 2013-02-19 21:43:37,462 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.7402,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 236.94
> > >> >> >>>> 2013-02-19 21:43:52,463 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.6961,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 251.941
> > >> >> >>>> 2013-02-19 21:44:07,464 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.657, MBytes/sec
> > >> >> received
> > >> >> >>>> =
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 266.942
> > >> >> >>>> 2013-02-19 21:44:37,465 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.5906,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 296.943
> > >> >> >>>> 2013-02-19 21:45:07,466 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.5364,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 326.944
> > >> >> >>>> 2013-02-19 21:45:22,467 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.5129,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 341.945
> > >> >> >>>> 2013-02-19 21:45:37,468 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.4913,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 356.946
> > >> >> >>>> 2013-02-19 21:46:07,469 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.4532,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 386.947
> > >> >> >>>> 2013-02-19 21:46:22,470 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.4363,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 401.948
> > >> >> >>>> 2013-02-19 21:46:52,471 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.406, MBytes/sec
> > >> >> received
> > >> >> >>>> =
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 431.949
> > >> >> >>>> 2013-02-19 21:47:07,472 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.3924,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 446.95
> > >> >> >>>> 2013-02-19 21:47:37,473 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.3677,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 476.951
> > >> >> >>>> 2013-02-19 21:48:07,474 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.346, MBytes/sec
> > >> >> received
> > >> >> >>>> =
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 506.952
> > >> >> >>>> 2013-02-19 21:48:22,475 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.336, MBytes/sec
> > >> >> received
> > >> >> >>>> =
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 521.953
> > >> >> >>>> 2013-02-19 21:48:52,476 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.3177,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 551.954
> > >> >> >>>> 2013-02-19 21:49:07,477 INFO org.apache.giraph.comm.netty.**
> > >> >> >>>> NettyClient:
> > >> >> >>>> waitSomeRequests: Waiting interval of 15000 msecs, 29 open
> > >> requests,
> > >> >> >>>> waiting for it to be <= 0, MBytes/sec sent = 0.3093,
> MBytes/sec
> > >> >> received
> > >> >> >>>>
> > >> >> >>> =
> > >> >> >>>
> > >> >> >>>> 0, MBytesSent = 175.3834, MBytesReceived = 0.0069, ave sent
> req
> > >> >> MBytes =
> > >> >> >>>> 0.2998, ave received req MBytes = 0, secs waited = 566.955
> > >> >> >>>>
> > >> >> >>>> It looks similar to the problem we had earlier with
> concurrency.
> > >> >> >>>>
> > >> >> >>>> Are you guys experiencing something similar at large-scale?
> > >> >> >>>>
> > >> >> >>>>
> > >> >> >>>
> > >> >> >>
> > >> >> >
> > >> >>
> > >> >>
> > >> >> --
> > >> >>    Claudio Martella
> > >> >>    [email protected]
> > >> >>
> > >> >
> > >> >
> > >>
> > >
> > >
> > >
> > >--
> > >   Claudio Martella
> > >   [email protected]
> >
> >
>



-- 
   Claudio Martella
   [email protected]

Reply via email to